COVID-19 Classification from Chest X-Ray Images: A Framework of Deep Explainable Artificial Intelligence.

Area of Science:

• Medical Image Analysis
• Artificial Intelligence in Healthcare
• Radiology

Background:

• Manual diagnosis of COVID-19 from chest X-rays is time-consuming and requires expert radiologists.
• Accurate and rapid COVID-19 detection is crucial for controlling its spread.
• Artificial intelligence (AI), particularly deep learning and explainable AI (XAI), offers promising solutions for medical image analysis.

Purpose of the Study:

• To propose and evaluate a novel deep learning and explainable AI technique for COVID-19 diagnosis and classification using chest X-ray images.
• To enhance the accuracy and efficiency of COVID-19 detection compared to manual methods.

Main Methods:

• A hybrid contrast enhancement technique was applied to chest X-ray images.
• Two modified deep learning models, utilizing deep transfer learning for feature extraction, were trained.
• Features were fused using improved canonical correlation analysis, optimized by the Whale-Elephant Herding algorithm, and classified with an extreme learning machine (ELM). Grad-CAM was used for visualization.

Main Results:

• The proposed method achieved high classification accuracies of 99.1%, 98.2%, and 96.7% on three public datasets.
• An ablation study confirmed the superiority of the proposed approach over existing methods.
• Explainable AI techniques (Grad-CAM) provided visual insights into the model's decisions.

Conclusions:

• The developed deep learning and explainable AI framework demonstrates high efficacy for automated COVID-19 detection and classification from chest X-rays.
• This approach offers a potential solution to overcome the limitations of manual radiological diagnosis, improving diagnostic speed and accuracy.
• The integration of explainable AI enhances the trustworthiness and interpretability of the automated diagnostic system.